Funding outcome

Gold standard treatment for cure in women with cervical cancer includes brachytherapy (placement of radioactive sources inside or near a tumor using applicators) to deliver high localized radiation doses. Safe use of such doses demands treatment accuracy through 'tailored' applicators and their precise placement, and delineation of organs for treatment planning. However, applicator insertion significantly deforms the internal anatomy, especially of the uterus. Currently, to pre-select appropriate applicators, physicians often use best judgment to predict this change using magnetic resonance images (MRI) taken without applicators to deliver safe, targeted high doses. Moreover, to minimize dose to healthy organs surrounding the tumor during brachytherapy, they are manually delineated in the planning software when designing the treatment. Better predictions are possible with machine learning, an artificial intelligence method, to detect patterns among data, learn from it, and make predictions, a process called predictive modeling. This project will apply machine learning in brachytherapy to predict applicator-related anatomic changes in 3-dimension (3D) and to automatically identify organs on MRI. Based on this prediction, customizable 3D printed applicators will be designed for individual patients after accounting for patient anatomy and tumor geometry. Predictive modeling offers promise to improve processes, by personalizing and automating key steps, on the path towards more consistent high-quality cervical brachytherapy. Such improvements can have major implications on patient outcomes, likely reducing morbidity and mortality among cervical cancer survivors.